Data Scientist Nanodegree

Supervised Learning

Project: Finding Donors for CharityML

Welcome to the first project of the Data Scientist Nanodegree! In this notebook, some template code has already been provided for you, and it will be your job to implement the additional functionality necessary to successfully complete this project. Sections that begin with 'Implementation' in the header indicate that the following block of code will require additional functionality which you must provide. Instructions will be provided for each section and the specifics of the implementation are marked in the code block with a 'TODO' statement. Please be sure to read the instructions carefully!

In addition to implementing code, there will be questions that you must answer which relate to the project and your implementation. Each section where you will answer a question is preceded by a 'Question X' header. Carefully read each question and provide thorough answers in the following text boxes that begin with 'Answer:'. Your project submission will be evaluated based on your answers to each of the questions and the implementation you provide.

Note: Please specify WHICH VERSION OF PYTHON you are using when submitting this notebook. Code and Markdown cells can be executed using the Shift + Enter keyboard shortcut. In addition, Markdown cells can be edited by typically double-clicking the cell to enter edit mode.

Getting Started

In this project, you will employ several supervised algorithms of your choice to accurately model individuals' income using data collected from the 1994 U.S. Census. You will then choose the best candidate algorithm from preliminary results and further optimize this algorithm to best model the data. Your goal with this implementation is to construct a model that accurately predicts whether an individual makes more than $50,000. This sort of task can arise in a non-profit setting, where organizations survive on donations. Understanding an individual's income can help a non-profit better understand how large of a donation to request, or whether or not they should reach out to begin with. While it can be difficult to determine an individual's general income bracket directly from public sources, we can (as we will see) infer this value from other publically available features.

The dataset for this project originates from the UCI Machine Learning Repository. The datset was donated by Ron Kohavi and Barry Becker, after being published in the article "Scaling Up the Accuracy of Naive-Bayes Classifiers: A Decision-Tree Hybrid". You can find the article by Ron Kohavi online. The data we investigate here consists of small changes to the original dataset, such as removing the 'fnlwgt' feature and records with missing or ill-formatted entries.


Exploring the Data

Run the code cell below to load necessary Python libraries and load the census data. Note that the last column from this dataset, 'income', will be our target label (whether an individual makes more than, or at most, $50,000 annually). All other columns are features about each individual in the census database.

Implementation: Data Exploration

A cursory investigation of the dataset will determine how many individuals fit into either group, and will tell us about the percentage of these individuals making more than \$50,000. In the code cell below, you will need to compute the following:

HINT: You may need to look at the table above to understand how the 'income' entries are formatted.

Featureset Exploration


Preparing the Data

Before data can be used as input for machine learning algorithms, it often must be cleaned, formatted, and restructured — this is typically known as preprocessing. Fortunately, for this dataset, there are no invalid or missing entries we must deal with, however, there are some qualities about certain features that must be adjusted. This preprocessing can help tremendously with the outcome and predictive power of nearly all learning algorithms.

Transforming Skewed Continuous Features

A dataset may sometimes contain at least one feature whose values tend to lie near a single number, but will also have a non-trivial number of vastly larger or smaller values than that single number. Algorithms can be sensitive to such distributions of values and can underperform if the range is not properly normalized. With the census dataset two features fit this description: 'capital-gain' and 'capital-loss'.

Run the code cell below to plot a histogram of these two features. Note the range of the values present and how they are distributed.

For highly-skewed feature distributions such as 'capital-gain' and 'capital-loss', it is common practice to apply a logarithmic transformation on the data so that the very large and very small values do not negatively affect the performance of a learning algorithm. Using a logarithmic transformation significantly reduces the range of values caused by outliers. Care must be taken when applying this transformation however: The logarithm of 0 is undefined, so we must translate the values by a small amount above 0 to apply the the logarithm successfully.

Run the code cell below to perform a transformation on the data and visualize the results. Again, note the range of values and how they are distributed.

Normalizing Numerical Features

In addition to performing transformations on features that are highly skewed, it is often good practice to perform some type of scaling on numerical features. Applying a scaling to the data does not change the shape of each feature's distribution (such as 'capital-gain' or 'capital-loss' above); however, normalization ensures that each feature is treated equally when applying supervised learners. Note that once scaling is applied, observing the data in its raw form will no longer have the same original meaning, as exampled below.

Run the code cell below to normalize each numerical feature. We will use sklearn.preprocessing.MinMaxScaler for this.

Implementation: Data Preprocessing

From the table in Exploring the Data above, we can see there are several features for each record that are non-numeric. Typically, learning algorithms expect input to be numeric, which requires that non-numeric features (called categorical variables) be converted. One popular way to convert categorical variables is by using the one-hot encoding scheme. One-hot encoding creates a "dummy" variable for each possible category of each non-numeric feature. For example, assume someFeature has three possible entries: A, B, or C. We then encode this feature into someFeature_A, someFeature_B and someFeature_C.

someFeature someFeature_A someFeature_B someFeature_C
0 B 0 1 0
1 C ----> one-hot encode ----> 0 0 1
2 A 1 0 0

Additionally, as with the non-numeric features, we need to convert the non-numeric target label, 'income' to numerical values for the learning algorithm to work. Since there are only two possible categories for this label ("<=50K" and ">50K"), we can avoid using one-hot encoding and simply encode these two categories as 0 and 1, respectively. In code cell below, you will need to implement the following:

Shuffle and Split Data

Now all categorical variables have been converted into numerical features, and all numerical features have been normalized. As always, we will now split the data (both features and their labels) into training and test sets. 80% of the data will be used for training and 20% for testing.

Run the code cell below to perform this split.


Evaluating Model Performance

In this section, we will investigate four different algorithms, and determine which is best at modeling the data. Three of these algorithms will be supervised learners of your choice, and the fourth algorithm is known as a naive predictor.

Metrics and the Naive Predictor

CharityML, equipped with their research, knows individuals that make more than \$50,000 are most likely to donate to their charity. Because of this, *CharityML* is particularly interested in predicting who makes more than \$50,000 accurately. It would seem that using accuracy as a metric for evaluating a particular model's performace would be appropriate. Additionally, identifying someone that does not make more than \$50,000 as someone who does would be detrimental to *CharityML*, since they are looking to find individuals willing to donate. Therefore, a model's ability to precisely predict those that make more than \$50,000 is more important than the model's ability to recall those individuals. We can use F-beta score as a metric that considers both precision and recall:

$$ F_{\beta} = (1 + \beta^2) \cdot \frac{precision \cdot recall}{\left( \beta^2 \cdot precision \right) + recall} $$

In particular, when $\beta = 0.5$, more emphasis is placed on precision. This is called the F$_{0.5}$ score (or F-score for simplicity).

Looking at the distribution of classes (those who make at most \$50,000, and those who make more), it's clear most individuals do not make more than \$50,000. This can greatly affect accuracy, since we could simply say "this person does not make more than \$50,000" and generally be right, without ever looking at the data! Making such a statement would be called naive, since we have not considered any information to substantiate the claim. It is always important to consider the naive prediction for your data, to help establish a benchmark for whether a model is performing well. That been said, using that prediction would be pointless: If we predicted all people made less than \$50,000, CharityML would identify no one as donors.

Note: Recap of accuracy, precision, recall

Accuracy measures how often the classifier makes the correct prediction. It’s the ratio of the number of correct predictions to the total number of predictions (the number of test data points).

Precision tells us what proportion of messages we classified as spam, actually were spam. It is a ratio of true positives(words classified as spam, and which are actually spam) to all positives(all words classified as spam, irrespective of whether that was the correct classificatio), in other words it is the ratio of

[True Positives/(True Positives + False Positives)]

Recall(sensitivity) tells us what proportion of messages that actually were spam were classified by us as spam. It is a ratio of true positives(words classified as spam, and which are actually spam) to all the words that were actually spam, in other words it is the ratio of

[True Positives/(True Positives + False Negatives)]

For classification problems that are skewed in their classification distributions like in our case, for example if we had a 100 text messages and only 2 were spam and the rest 98 weren't, accuracy by itself is not a very good metric. We could classify 90 messages as not spam(including the 2 that were spam but we classify them as not spam, hence they would be false negatives) and 10 as spam(all 10 false positives) and still get a reasonably good accuracy score. For such cases, precision and recall come in very handy. These two metrics can be combined to get the F1 score, which is weighted average(harmonic mean) of the precision and recall scores. This score can range from 0 to 1, with 1 being the best possible F1 score(we take the harmonic mean as we are dealing with ratios).

Question 1 - Naive Predictor Performace

Please note that the the purpose of generating a naive predictor is simply to show what a base model without any intelligence would look like. In the real world, ideally your base model would be either the results of a previous model or could be based on a research paper upon which you are looking to improve. When there is no benchmark model set, getting a result better than random choice is a place you could start from.

HINT:

Supervised Learning Models

The following are some of the supervised learning models that are currently available in scikit-learn that you may choose from:

Question 2 - Model Application

List three of the supervised learning models above that are appropriate for this problem that you will test on the census data. For each model chosen

HINT:

Structure your answer in the same format as above^, with 4 parts for each of the three models you pick. Please include references with your answer.

Answer:

  1. Support Vector Machines (SVM)
    • Real World Application: SVMs can be used in Bioinformatics to classifying protiens, genses, and cancer cells
    • Strengths: SVMs are shine when it comes to high dimensionality. It performs well when the margin of separation between classes is clear
    • Weaknesses: SVMs may be susceptible to noise in the dataset and underperforms with large datasets. It may perform poorly when target classes overlap.
    • Selection Reason: SVMs can support the large number of features included in the dataset to to provide accurate results.
  1. K-Nearest Neighbors (KNeighbors)
    • Real World Application: KNNs can be implemented in the finance industry for loan management and stock market forecasting
    • Strengths: KNNs are simple to implement and do not require a training period. They permorm well when adding new data, in which this addition will not affect the accuracy of the algorithm.
    • Weaknesses: KNNs are sensitive to outliers and missing values. They perform poorly when features are not homogenuous or scaled.
    • Selection Reason: KNNs require no training period in comparison to SVMs or Random Forests. They will also provide high accuracy when it comes to our clearly defined binary classification (>50k or <=50k)
  1. Ensemble Methods (Random Forest)
    • Real World Application:
    • Strengths: Random forests are robust to noisy, non-linear data. Since they're bagging models, Radnom forests are parellizable which results in faster computation time. They perform well with a mixture of numerical and categorical features
    • Weaknesses: Random forests provide slow real time prediction. They are difficult to implement due to complex algorithms. Random forests may work poorly for very large datasets since the size of the trees can take up a lot of memory.
    • Selection Reason: Random forests work well with high dimensional data since we are working with subsets of data. Our dataset has a lot of features

References (end of notebook): [13] [14] [15] [21] [22] [23] [24] [25] [26] [27] [28]

Implementation - Creating a Training and Predicting Pipeline

To properly evaluate the performance of each model you've chosen, it's important that you create a training and predicting pipeline that allows you to quickly and effectively train models using various sizes of training data and perform predictions on the testing data. Your implementation here will be used in the following section. In the code block below, you will need to implement the following:

Implementation: Initial Model Evaluation

In the code cell, you will need to implement the following:

Note: Depending on which algorithms you chose, the following implementation may take some time to run!


Improving Results

In this final section, you will choose from the three supervised learning models the best model to use on the student data. You will then perform a grid search optimization for the model over the entire training set (X_train and y_train) by tuning at least one parameter to improve upon the untuned model's F-score.

Question 3 - Choosing the Best Model

HINT: Look at the graph at the bottom left from the cell above(the visualization created by vs.evaluate(results, accuracy, fscore)) and check the F score for the testing set when 100% of the training set is used. Which model has the highest score? Your answer should include discussion of the:

Answer:

The three models chosen to test were SVM, KNN, and Random Forest. The visual analysis of each performance is shown above. It can be noticed that all models perform considerably better when it come to the training dataset, which is expected. However, the performance on the testing dataset is what is of importance to us.

Based on the F-Score for 100% training set, both the random forest classifier and SVM fall into the higher range in comparison to KNN, approximately 65%. Therefore, with regards to performance, SVM and Random Forest provided better metrics. However, SVM exhibited very high prediction/training time thus providing higher expected computational cost in comparison to Random Forest. In addition, the census dataset is suited well with RFs since they work well with impbalanced skewed data.

All in all, for the task of identifying individuals that make more than $50,000, the chosen model is Random Forest Classifiers as per the discussion above.

Question 4 - Describing the Model in Layman's Terms

HINT:

When explaining your model, if using external resources please include all citations.

Answer:

Random Forest is a machine learning algorithm, which in our case we will use to predict potential donors based on income. Our main target for this project is to identify donors by identifying individual who make more than $50k.

The basic building blocks of RFs are decision trees, in which RF merges them together to get an accurate and stable prediction. RFs think of these smaller DTs as weak classifiers, and it combine these weak classifiers (DT) to obtain a strong classifier (RF). To further understand RFs, we need to understand decision trees. An example of a decision tree is shown below. Decision Tree A decision tree, as the name implies, is a tree-like structure which reaches a decision or predicition by answering questions 9conditions). In the example above, the model predicts if a person is fit by asking about their age and lifestyle.

RFs follow the divide and conquer approach, in which it divides the data into smaller set and each set has a corresponding decision tree with predictor variables (age, education, marital status, occupation, race, etc...). Each DT makes predictions depending on different variables. What RF does is it take the prediction from all DTs under it, and in our case it takes the majority vote (>50k or <=50k).
References (end of notebook): [29] [30] [31]

Implementation: Model Tuning

Fine tune the chosen model. Use grid search (GridSearchCV) with at least one important parameter tuned with at least 3 different values. You will need to use the entire training set for this. In the code cell below, you will need to implement the following:

Note: Depending on the algorithm chosen and the parameter list, the following implementation may take some time to run!

Question 5 - Final Model Evaluation

Note: Fill in the table below with your results, and then provide discussion in the Answer box.

Results:

Metric Unoptimized Model Optimized Model
Accuracy Score 0.8413 0.8583
F-score 0.6789 0.7315

Answer:

Shown in the table above are the optimized model's accuracy and F-score on the testing data in comparison to that of the unoptimized model. It is noticed that the F-Score has increased by ~5% and the accuracy score by ~1.5% after optimization and tuning the hyperparameters.
Furthermore, in comparison to the naive predictor, the F-Score has increased by ~45% and the accuracy score by ~60%, which is a very notable but expected improvement.


Feature Importance

An important task when performing supervised learning on a dataset like the census data we study here is determining which features provide the most predictive power. By focusing on the relationship between only a few crucial features and the target label we simplify our understanding of the phenomenon, which is most always a useful thing to do. In the case of this project, that means we wish to identify a small number of features that most strongly predict whether an individual makes at most or more than \$50,000.

Choose a scikit-learn classifier (e.g., adaboost, random forests) that has a feature_importance_ attribute, which is a function that ranks the importance of features according to the chosen classifier. In the next python cell fit this classifier to training set and use this attribute to determine the top 5 most important features for the census dataset.

Question 6 - Feature Relevance Observation

When Exploring the Data, it was shown there are thirteen available features for each individual on record in the census data. Of these thirteen records, which five features do you believe to be most important for prediction, and in what order would you rank them and why?

Answer:

Out of the 13 features I believe the following features affect the prediction most (ranked by most important):

  1. capital-gain
  2. occupation
  3. marital-status
  4. age
  5. education

I believe capital gain is of utmost importance since it provides some insight into the financial aspect of an individual's life. After that we have occupation as salaries depend on occupation. For the third feature, I chose marital status as it plays a role if a couple share financial assets. Fourth, age affects the donor prediction as individuals making >50k tend to be older one might say. Lastly, an educated individual is more like to donate and have a higher income than $50k.

Implementation - Extracting Feature Importance

Choose a scikit-learn supervised learning algorithm that has a feature_importance_ attribute availble for it. This attribute is a function that ranks the importance of each feature when making predictions based on the chosen algorithm.

In the code cell below, you will need to implement the following:

Question 7 - Extracting Feature Importance

Observe the visualization created above which displays the five most relevant features for predicting if an individual makes at most or above \$50,000.

Answer:

In comparison to the guesses I made, four out of 5 guesses were correct. The feature that was not on my list was the relationship, which in hindsight does in fact play a role since husbands tend to earn more compared to their wifes. The ranking is also somewhat similar to the visulization, in which the visualization also ranked capital gain with the most weight. However, it ranked education higher than age contrary to my rankings. This may be due to the fact that both features are somewhat related. In other words, one may infer a person's age as per their education. And looking at the bigger picture, it preferable to know one's educational level rather than age for predicting their income.

Feature Selection

How does a model perform if we only use a subset of all the available features in the data? With less features required to train, the expectation is that training and prediction time is much lower — at the cost of performance metrics. From the visualization above, we see that the top five most important features contribute more than half of the importance of all features present in the data. This hints that we can attempt to reduce the feature space and simplify the information required for the model to learn. The code cell below will use the same optimized model you found earlier, and train it on the same training set with only the top five important features.

Question 8 - Effects of Feature Selection

Answer:

The table below shows the model's performance when the full data and reduced data were used.

Metric Full Data Reduced Data
Accuracy Score 0.8583 0.8457
F-score 0.7315 0.6933

It can be noticed that the using reduced data very mildly affected the F-score and accuracy, in which F-score decreased by 3.82% and accuracy decreased by 1.26%. Furthermore, if training time was a factor, I would in fact use the reduced dataset as it will decrease training time in addition to cut the features from 103 to 5, which will in turn decrease computational cost.

References

[1] https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.loc.html
[2] https://www.w3schools.com/python/ref_func_round.asp#:~:text=The%20round()%20function%20returns,will%20return%20the%20nearest%20integer
[3] https://stackoverflow.com/questions/53689432/using-pandas-map-to-change-values
[4] https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.get_dummies.html
[5] https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.train_test_split.html
[6] https://scikit-learn.org/stable/modules/generated/sklearn.metrics.fbeta_score.html
[7] https://scikit-learn.org/stable/modules/generated/sklearn.metrics.accuracy_score.html
[8] https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html#sklearn.ensemble.RandomForestClassifier
[9] https://scikit-learn.org/stable/modules/generated/sklearn.metrics.make_scorer.html
[10] https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.GridSearchCV.html
[11] https://scikit-learn.org/stable/modules/generated/sklearn.svm.SVC.html#sklearn.svm.SVC
[12] https://scikit-learn.org/stable/modules/generated/sklearn.neighbors.KNeighborsClassifier.html
[13] https://www.dataquest.io/blog/top-10-machine-learning-algorithms-for-beginners/
[14] https://developer.ibm.com/technologies/artificial-intelligence/articles/cc-supervised-learning-models/
[15] https://analyticsindiamag.com/7-types-classification-algorithms/
[16] https://stackoverflow.com/questions/36869258/how-to-use-graphviz-with-anaconda-spyder
[17] https://www.geeksforgeeks.org/python-os-system-method/
[18] https://towardsdatascience.com/how-to-visualize-a-decision-tree-from-a-random-forest-in-python-using-scikit-learn-38ad2d75f21c
[19] https://scikit-learn.org/stable/modules/generated/sklearn.tree.export_graphviz.html
[20] https://wordpress.com/support/markdown-quick-reference/?aff=13200
[21] https://medium.com/@dhiraj8899/top-4-advantages-and-disadvantages-of-support-vector-machine-or-svm-a3c06a2b107
[22] https://data-flair.training/blogs/svm-support-vector-machine-tutorial/
[23] http://theprofessionalspoint.blogspot.com/2019/02/advantages-and-disadvantages-of-knn.html
[24] https://www.fromthegenesis.com/pros-and-cons-of-k-nearest-neighbors/
[25] https://www.ijera.com/papers/Vol3_issue5/DI35605610.pdf
[26] https://datascience.stackexchange.com/questions/6838/when-to-use-random-forest-over-svm-and-vice-versa#:~:text=Random%20Forest%20is%20intrinsically%20suited,of%20numerical%20and%20categorical%20features
[27] http://theprofessionalspoint.blogspot.com/2019/02/advantages-and-disadvantages-of-random.html#:~:text=Random%20Forest%20is%20based%20on,and%20therefore%20improves%20the%20accuracy
[28] https://towardsdatascience.com/why-random-forest-is-my-favorite-machine-learning-model-b97651fa3706
[29] https://medium.com/@chiragsehra42/decision-trees-explained-easily-28f23241248
[30] https://builtin.com/data-science/random-forest-algorithm
[31] https://victorzhou.com/blog/intro-to-random-forests/

Note: Once you have completed all of the code implementations and successfully answered each question above, you may finalize your work by exporting the iPython Notebook as an HTML document. You can do this by using the menu above and navigating to
File -> Download as -> HTML (.html). Include the finished document along with this notebook as your submission.